LETTER

pubs.acs.org/NanoLett

Batteries for Efficient Energy Extraction from a

Water Salinity Difference
Fabio La Mantia,||,† Mauro Pasta,||,†,‡ Heather D. Deshazer,† Bruce E. Logan,§ and Yi Cui*,†
†
Department of Materials Science and Engineering, Stanford University, Stanford, California 94305, United States
‡
Dipartimento di Chimica Inorganica, Metallorganica e Analitica “Lamberto Malatesta”, Universita degli Studi di Milano, Via Venezian
21, 20133 Milan, Italy
§
Department of Civil and Environmental Engineering, Penn State University, University Park, Pennsylvania 16802, United States
bS Supporting Information
ABSTRACT: The salinity difference between seawater and river water is a renewable
source of enormous entropic energy, but extracting it efficiently as a form of useful
energy remains a challenge. Here we demonstrate a device called “mixing entropy
battery”, which can extract and store it as useful electrochemical energy. The battery,
containing a Na2
xMn5O10 nanorod electrode, was shown to extract energy from real
seawater and river water and can be applied to a variety of salt waters. We demonstrated
energy extraction efficiencies of up to 74%. Considering the flow rate of river water into
oceans as the limiting factor, the renewable energy production could potentially reach
2 TW, or ∼13% of the current world energy consumption. The mixing entropy battery is
simple to fabricate and could contribute significantly to renewable energy in the future.
KEYWORDS: Salinity-gradient power, mixing entropy, sodium intercalation, energy
harvesting

T echnologies to utilize renewable energy such as solar, wind,

geothermal, and biomass sources have attracted great atten-
tion recently. The large scale chemical energy stored as the salinity
difference between seawater and freshwater is another renewable
source which can be harvested. The major components of the
global water cycle involve distillation of water from oceans by
evaporation, precipitation, and collection of this freshwater in
rivers, lakes, and aquifers, with mixing of freshwater and salt water
in estuaries. Solar energy drives this cycle, creating a significant
salinity difference between seawater and freshwater. The entropic
energy created by the difference in water salinities is normally
dissipated when river water flows into the sea. This reduction in
free energy due to the mixing is estimated at 2.2 kJ of free energy
per liter of freshwater.
To date this significant and completely renewable energy source
has not been fully harnessed, although since Pattle’s pioneering
studies in 19541 several types of technologies have been proposed
in order to take advantage of this renewable energy source. Past
suggestions for capturing energy from the mixing of seawater and
freshwater include: hydroelectric pile,1 based on acidic and basic
membranes; pressure-retarded osmosis,2 based on semipermeable
membranes; reverse electrodialysis,3 based on ion selective mem-
branes; concentration electrochemical cells;4 and devices exploit-
ing differences in vapor pressures.5 Nowadays, the most reliable
methods use membranes as separator between the seawater and
the freshwater, to prevent the mixing of the two streams. There-
fore, the application of these technologies is strongly based on the
energy efficiency, cost, and lifetime of the membranes themselves.6
Recently, Brogioli has proposed a new and interesting method
based on the electrochemical double layer capacitor technology,
using activated carbon electrodes.7 This technology has intrinsic
technical challenges, like high sensitivity to impurities and
dissolved oxygen, causing self-discharge phenomena and the
use of the electrode/electrolyte interface for energy storage.
Sales et al. employed a membrane-modified electrochemical
double layer capacitor, improving the energy conversion effi-
ciency of the Brogioli system.8 Finally, Guo et al. have proposed a
technology based on single ion selective nanopores for powering
biomedical devices through salinity difference in body fluids.9
In this work, we demonstrate a novel electrochemical cell
named a “mixing entropy battery”, which extracts energy from
the difference in concentration of two solutions and stores it as
chemical energy inside the electrode material’s bulk crystal
structure. This approach allows us to overcome the challenges
of supercapacitor electrodes based on activated carbon. This
device consists of a reversible electrochemical system where the
salts in the electrolyte are the reactants and the electrode stores
ions. We employed two different electrodes: an anionic electrode,
which interacts with Cl
ions selectively; and a cationic electrode,
which interacts with Naþ ions selectively. These electrodes are
initially submerged in a low ionic strength solution (river water)
in their discharged states, when the electrode materials contain
the respective ions incorporated in their structures. In this dilute
Received: February 11, 2011

r XXXX American Chemical Society A dx.doi.org/10.1021/nl200500s | Nano Lett. XXXX, XXX, 000–000
Nano Letters LETTER

Figure 1. (a) Schematic representation of the working principle behind a complete cycle of the mixing entropy battery, showing how energy extraction
can be accomplished: step 1, charge in river water; step 2, exchange to seawater; step 3, discharge in seawater; step 4, exchange to river water. (b) Typical
form of a cycle of battery cell voltage (ΔE) vs charge (q) in a mixing entropy battery, demonstrating the extractable energy.

solution, the battery is charged by removing the Naþ and Cl

ions from the respective electrodes (Figure 1a, step 1). Succes-
sively, the dilute electrolyte is exchanged for a concentrated
solution (seawater), which is accompanied by an increase in the
potential difference between the electrodes (Figure 1a, step 2).
At this higher potential difference, the battery is discharged, as
the anions and cations are reincorporated into their respective
electrodes (Figure 1a, step 3). The concentrated solution is then
removed and substituted by the dilute electrolyte (river water),
which results in a decrease in potential difference between the
electrodes (Figure 1a, step 4). We note that the exchange of
solution could also be carried out via a flow process, which could
be attractive for large scale energy extraction.
This closed cycle produces energy, as shown by a schematic of
the expected shape of the battery cell voltage (ΔE) versus charge
(q) during one cycle in Figure 1a. We observe that during steps 2
and 4 no energy is produced or consumed. During step 1, the
battery requires energy to drive the ions out of the crystal structure,
while during step 3 the battery produces energy by incorporation
of the ions. Therefore, the energy gain is due to the fact that the
same amount of charge is released in step 3 at a higher voltage than
consumed in step 1. The energy gain is given by the integral along
the cycle of the voltage with respect to the charge Figure 2. Gibbs free energy of mixing as a function of the volumetric
I fraction of seawater (χ), calculated at 298 K, mixing a 0.6 M solution and
W ¼
ΔE dq ð1Þ a 0.024 M solution for per unit volume of (a) final solution and (a) low
C concentration solution. (c) Estimated extractable power (GW) from the
difference of salinity in different countries, based on flow of river water
The difference in the Gibbs free energy per unit of volume emptying into the ocean.
(ΔG) between the solution after mixing seawater and river water
and the separate seawater and river water before mixing can be
calculated as a function of the volumetric fraction of seawater, χ. range of salt concentrations examined here (0.024 M for river
The minimum in Figure 2a, which is the maxima of mixing water, and 0.60 M for seawater; see Supporting Information).
entropy of the two solutions, is obtained at χ ≈ 0.4 and is equal to In order to calculate the maximum extractable energy per unit
ΔG ≈
0.85 kJ dm
3 with respect to the mixed solution. volume of river water, which is the limiting reagent in this
Figure 2a is based on the equation process, the Gibbs free energy of mixing is plotted with respect
to the volume of river water consumed in Figure 2b. The energy
ΔG needed to mechanically exchange river water and seawater in the
¼ cT lnðcT Þ
χc1 lnðc1 Þ
ð1
χÞc2 lnðc2 Þ ð2Þ device is assumed to be negligible compared to the total energy
RT produced. For an excess of seawater reagent, the maximum
energy extractable is ΔG ≈
2.5 kJ dm
3 of river water
where R is the universal gas constant, T the absolute temperature, (obtained for χ ≈ 1). This means that a power plant processing
cT the concentration of the mixed solution, c1 the concentration of 40 m3 s
1 of river water could produce up to 100 MW.
the first solution, and c2 the concentration of the second solution. To demonstrate experimentally the validity of our mixing
The activity coefficient can be considered to be unity over the entropy battery design, we constructed an electrochemical cell
B dx.doi.org/10.1021/nl200500s |Nano Lett. XXXX, XXX, 000–000
Nano Letters
Figure 3. (a) Measured open circuit potentials for the AgCl/Na2M-
n5O10 system at different concentrations of NaCl in aqueous solution.
(b) Measured potential (E) of individual Ag/AgCl (red) and Na2
xM-
n5O10 (blue) electrodes vs Ag/AgCl reference electrode during the cycle
of energy extraction. (c) Energy extraction cycle for the AgCl/Na2M-
n5O10 system in a ΔE vs q plot (the area is equal to the extracted energy).
(d) Energy per area extracted vs number of cycles for the AgCl \/Na2
Mn5O10 system.
employing the following reaction
5MnO2 þ 2Ag þ 2NaCl T Na2 Mn5 O10 þ 2AgCl ðR1Þ
where the Na2
xMn5O10 electrode made of nanorods was used
to capture Naþ ions and the Ag electrode to capture Cl
ions.
(See Supporting Information for detailed electrode preparation.)
Among other Naþ intercalation materials, we selected Na2M-
n5O10 because of its energy density, low cost, and benign
environmental impact. The Ag/AgCl electrode was used because
it was readily available, although it would not be ideal for capturing
Cl
ions in practice since silver is a precious metal. The equilibrium
potential difference of the reaction was measured as a function of
the concentration of NaCl in the electrolyte (Figure 3a). As
expected, the equilibrium potential difference increased with NaCl
concentration (see Supporting Information for further details).
The gain in energy was correlated to the change in the potential
difference obtained at different NaCl concentrations. We refer to
the change in potential difference as the gained potential, VG. In
the case of seawater (0.6 M NaCl solution) and river water (0.024 M
NaCl solution), the thermodynamic gained potential was VG ≈
0.135 V (see Supporting Information). This value corresponds to
about 3
4 times the gained potential obtained using a super-
capacitor material (activated carbon) for the electrodes (VG ≈
0.033V6, VG ≈ 0.040V8).
In Figure 3b the typical cycle of the Ag/AgCl and Na2Mn5O10
system is reported for the two separate electrodes (vs Ag/AgCl/
KCl 3.5 M reference electrode). In the presence of the dilute
electrolyte, the potential difference of the two electrodes is small.
In step 1 the battery is charged by removing the ions from AgCl
and Na2Mn5O10, thus gradually increasing the potential differ-
ence of the two electrodes. The current density applied is
250 μA cm
2, and a time limit of 20 min is imposed (see Supporting
Information). During step 2, the dilute solution is substituted
with the concentrated one, and consequently the potential
C dx.doi.org/10.1021/nl200500s |Nano Lett. XXXX, XXX, 000–000
LETTER
difference between the two electrodes increases significantly.
During step 3, the battery is discharged, and the ions are captured
by the electrodes, while the potential difference gradually de-
creases. The applied current density is
250 μA cm
2, with a time
limit of 20 min. The voltage difference between the Ag/AgCl and
Na2Mn5O10 electrodes is plotted during this cycling process (ΔE
vs q) in Figure 3c. After one full cycle, the energy density produced
is about 29 mJ cm
2 (power density 10.5 μW cm
2).
The spikes in the potential difference observed at the begin-
ning and end of step 1 (A and A0 ) and step 3 (B and B0 ) are due to
electrochemical losses, known as overpotential. Due to this loss,
the practical gained potential is VG ≈ 0.100 V (74% efficiency, as
compared to the thermodynamically predicted potential gain). It
is possible to optimize the experimental configuration to improve
this gain as well, as the electrodes are currently at a distance of
approximately 1 cm from each other. The resistance of the
solution was 75 Ω for the diluted solution (75% of the total
internal resistance of the device) and 5 Ω in the concentrated one
(20% of the total internal resistance of the device), creating a high
overpotential. This overpotential could be greatly decreased (and
thus the power and/or energy efficiency significantly increased)
by designing a more optimized cell geometry using closer
electrode spacing and allowing a small amount of seawater to
remain in the cell when adding the river water to reduce the
resistance of the solution (see Supporting Information). Thus,
this system can be further optimized in order to increase the
energy recovery in each cycle. The experimental energy pro-
duced with respect to cycle number is reported in Figure 3d. It is
important to note that the system operation was very stable for
100 cycles, with essentially no observable loss in energy produc-
tion over this period. This consistent power production is
expected, since both anionic and cationic electrodes are always
operating within the stability window of water, and the electrode
materials are very stable in the aqueous environment.
In order to probe possible complications which could exist in a
real system, we also conducted measurements with real water
samples collected from local natural water sources (see Support-
ing Information). There was no decrease in cycling performance
when these samples were used, and we observed no electrode
degradation, self-discharge, or other detrimental phenomenon.
The energy conversion efficiency of this cell was 75%.
The extractable power from the difference of salinity of river
and seawater that could be obtained in many different countries
around the world is summarized in Figure 2c (data from ref 10).
If the energy from mixing entropy was harnessed from all these
rivers, the power obtained could reach up to 2 TW, which is
∼13% of the current total global energy requirement.11 This
energy can also be easily harvested at low temperatures, and is
completely renewable, since the ultimate source is the solar
energy which powers the water cycle.
Thus far, we have only considered the possibility of energy
production by harvesting the free energy dissipated by mixing
seawater and river water. However, this system could be modified
to operate on a smaller scale using solar energy to distill water,
with complete recycle of the electrolyte to achieve conversion of
solar energy into electrical energy. To further examine this
concept, we investigated electrolytes based on the reaction
FePO4 þ Ag þ LiCl T LiFePO4 þ AgCl ðR2Þ
LiCl is highly soluble (832 g dm
3 in water) and LiFePO4/FePO4
is a two-phase system with a well defined potential capable of rapid
Nano Letters LETTER

Author Contributions
F.L., M.P., and Y.C. conceived and designed the experiments. F.
L., M.P., and H.D.D. performed the experiments. All authors
contributed to the discussion and writing of the paper.

’ ACKNOWLEDGMENT
The work is partially supported by King Abdullah University
of Science and Technology (KAUST) under awards KUS-l1-
001-12 and KUS-I1-003-13 and by the Department of Energy,
Office of Basic Energy Sciences, Division of Materials Sciences
and Engineering, under contract DE-AC02-76SF0051, through
the SLAC National Accelerator Laboratory LDRD project.

’ REFERENCES
(1) Pattle, R. E. Nature (London, U.K.) 1954, 174, 660.
(2) Levenspiel, O.; De Nevers, N. Science 1974, 183 (4121), 157–60.
Figure 4. (a) Measured open circuit potential for the AgCl/LiFePO4 (3) Weinstein, J. N.; Leitz, F. B. Science 1976, 191 (4227), 557–9.
system at different concentrations of NaCl in aqueous solution. (b) (4) Clampitt, B. H. K.; F., E. Science 1976, 194, 719.
Measured potential of individual Ag/AgCl (red) and FePO4/LiFePO4 (5) Olsson, M. W.; G., L; Isaacs, J. D. Science 1979, 189, 654.
(blue) electrodes vs Ag/AgCl reference electrode during the cycle of (6) Jones, A. T.; Finley, W. Oceans 2003. Proceedings DOI: 10.1109/
energy extraction. (c) Energy extraction cycle for the AgCl/LiFePO4 OCEANS.2003.1282847.
system in a ΔE vs q plot (the area is equal to the extracted energy). (d) (7) Brogioli, D. Phys. Rev. Lett. 2009, 103 (5), 058501/1–058501/4.
Energy extracted vs number of cycles for the AgCl/LiFePO4 system. (8) Sales, B. B.; Saakes, M.; Post, J. W.; Buisman, C. J. N.; Biesheuvel,
P. M.; Hamelers, H. V. M. Environ. Sci. Technol. 2010, 44 (14),
5661–5665.
charging and discharging. With the exception of Ag, these
(9) Guo, W.; Cao, L.; Xia, J.; Nie, F.-Q.; Ma, W.; Xue, J.; Song, Y.;
materials are inexpensive and environmentally benign. In tests Zhu, D.; Wang, Y.; Jiang, L. Adv. Funct. Mater. 2010, 20 (8), 1339–1344.
with 0.03 and 1.5 M solutions, we produce an energy density of (10) Data from “The global river discharge database”.
38.2 mJ cm
2 (power density 13.8 μW cm
2) that was stable over (11) Moriarty, P.; Honnery, D. Int. J. Hydrogen Energy 2009, 34 (1),
100 cycles (Figure 4). These results indicate that this system has 31–39.
significant potential as a method of harnessing solar energy, and
has the possibility to be completely renewable, due to repeated
cycling of the materials with no loss of reagent.
The mixing entropy battery and other processes described
here represent not only a novel type of electrochemical system
compared to existing technologies but also a very promising
technology whose practical application can make a significant
contribution to the field of renewable energy production. This
process for generating electrical energy (Figure 1a) could also be
reversed, and exploited as a method for water desalination.
Future work will focus on cell geometry optimization and the
development of materials which can be used to replace silver as
the anionic electrode.

’ ASSOCIATED CONTENT

bS Supporting Information. Additional details on electro-

chemical thermodynamics details, synthesis and characterization
of the Naþ capturing electrodes, electrode preparation and
electrochemical characterization, and tests with collected water
samples and additional figures. This material is available free of
charge via the Internet at http://pubs.acs.org.

’ AUTHOR INFORMATION
Corresponding Author
*E-mail: yicui@stanford.edu.
Author Contributions
)

These authors contributed equally to this work.

D dx.doi.org/10.1021/nl200500s |Nano Lett. XXXX, XXX, 000–000